More modelling recipes

With (life) insurance data

Pat Reen

Overview

Background

This document sets out a few practical recipes for modelling with (life) insurance data. Insurance events are typically of low probability - these recipes consider some of the limitations of “small data” model fitting (where the observations of interest are sparse) and other topics for insurance like comparisons to standard tables. Themes

  • Common data transforms, summary stats, and simple visualisations
  • Regression
    • Grouped vs ungrouped data
    • Choice of: response distribution, link (and offsets), explanatory variables
    • Modelling variance to industry/ reference (A/E or A - E)
    • Model selection: stepwise regression, likelihood tests, model evaluation
    • Predictions, confidence intervals and visualisations
  • Bayesian regression and other classification models - to follow.

See link above to GitHub repository which has the detailed code.